云计算环境下大规模Web服务故障诊断技术研究

Research on Fault Diagnosis for Massive Web Services in Cloud Computing

摘要:

随着Web服务规模的不断增大,系统中往往会伴随着大量的噪音数据；为了提高Web服务诊断在噪音影响下的健壮性,文章研究了大规模Web服务下的故障诊断；首先,提出了一种Web服务自动诊断框架；其次,分别定义了转移矩阵和发射矩阵,并通过这两种矩阵来表达Web服务的状态转换过程；最后,提出了一种基于Viterbi方法的故障诊断算法；实验表明,文章提出的算法与其它相关的Web服务故障诊断算法相比准确性高,并且对噪音数据有着良好的健壮性。

Abstract:

As the scale of Web services is becoming larger, there are always noise data in the system. In order to improve the robustness of fault diagnosis of the Web services, this paper studies the problem of fault diagnosis for Web services. Firstly, we proposed an automatic diagnosis framework for Web service. Secondly, we defined the transition and emission matrixes, respectively, and used these matrixes to represent the state transition of Web services. Finally, we proposed a Viterbi based fault diagnosis algorithm. The experiments show that, the proposed algorithm has higher accuracy than related works, and has good robustness for noise data in Web services.